Statistics for the Behavioral and Social Sciences

Reduction of Prejudice

The Contact Hypothesis of Gordon Allport and the Reduction of Prejudice

The literature covering the nature of prejudice, its scope, the effects of prejudice, and methods to reduce on prejudice is among the most extraordinary body of literature in all of social science. The total volume of research on the topic of prejudice is quite extraordinary and this body of work reflects several decades of scholarly investigation of the meaning of prejudice, its assessment, its etiology, its consequences, and methods to reduce prejudice. There are very few areas of study that have attracted a greater range of theoretical perspectives than the area of prejudice. Theorizing about the nature and manifestation of prejudice has also been accompanied by many spirited debates about the appropriate way to conceptualize methods to reduce prejudice in people. The result has been a rich body of measurement instruments and reduction strategies. The most enduring of the strategies to reduce prejudice is known as the contact hypothesis, which states that under certain conditions contact between groups leads to a decrease in prejudice. The current paper reviews some of the literature for and against the effectiveness of contact.

Statement of the Problem

For the purposes of this paper we will describe prejudice as it is normally defined in the literature as an unfair negative attitude directed towards a particular social group or directed towards a member of that particular social group. Stereotypes, on the other hand are defined as overgeneralizations about a particular group that in effect are especially rigid, most often are factually incorrect, and this set of beliefs often goes along with the negative feelings and attitudes that are associated with prejudice. Discrimination is defined as the unjust treatment of members of these different social groups originating from prejudicial attitudes and stereotyping (Allport, 1954). Thus, prejudicial attitudes toward certain groups are the driving force behind certain negative beliefs and unscrupulous behaviors directed towards certain groups. Developing functional and practical methods to reduce prejudice should lead to reductions in stereotyping and discrimination.

While the methodological complexities and the abundance of theoretical contributions provided by the literature on prejudice are impressive, it can sometimes be less clear with regards to applying the practical knowledge that it all of this research and theorizing has generated. The study of prejudice creates a center of attention because researchers and other scholars endeavor to understand and cure the social problems associated with it such as inequality, discrimination, and prejudicially motivated violence. These goals are also shared by many government officials and huge sums of money are spent annually on prospective interventions aimed at reducing prejudice in our schools, in the workplace, and in areas plagued with intergroup conflict. Given these practical objectives associated with the research on prejudice it is important to understand what the extensive body of research has learned regarding some of the more effective ways to reduce prejudice. The current paper focuses on the most enduring theoretical method to reduce prejudice in people, Gordon Allport's contact hypothesis (Allport, 1954).

Literature Review

Strategies for reducing prejudice that rely on an intergroup approach originate from early theories that the perception and actions of most people are designed to be favorable towards members of their own group (termed "the ingroup" in the literature) relative to groups that they are not a member of (termed "the outgroup"). There have traditionally been two major theoretical lines of thinking that have inspired specific techniques to deal with this in-group/out-group bias in people. The first of these is Gordon Allport's contact hypothesis (Allport 1954), which states that exposure to members of the out-group under certain optimal conditions will result in a reduction in prejudice. The second group of theories are termed "social identity and categorization theories" (Tajfel & Turner, 1979). These theories advocate interventions that break down or rearrange social boundaries. However the contact hypothesis has received far more attention in the literature.

The contact hypothesis of Gordon Allport states that under certain types of favorable conditions contact between two groups should lead to a significant reduction in prejudice. According to Allport there are four conditions of contact that facilitate positive attitudes and reduce prejudice: groups must have equal status within the contact situation, there should be no competition within the contact situation, the groups must seek superordinate goals together within the contact situation, and the authorities must approve the contact and strongly endorse a reduction in tensions between the groups.

The classic study cited in undergraduate and graduate texts to provide evidence for the validity of the contact hypothesis is the "Robbers Cave" study (Sherif, Harvey, White, Hood, & Sherif, 1961). The subjects of this experiment were 22 eleven-year-old boys. These boys were taken to a summer camp in Robbers Cave State Park which is in the state of Oklahoma. Prior to the actual experiment, the subjects were randomly divided into two groups. The groups were then separately transported to the park and housed in cabins. Initially the two groups were not aware of the existence of the other group. During the first week of the experiment each group spent time bonding with each other. The groups also developed a name for themselves which was put on flags in the camp and on their shirts. The groups were named the Eagles and Rattlers respectively.

Eventually, the groups learned of the existence of the other group. Then a series of competitive activities between the groups was performed which resulted in rewards for the members of the winning group. As the antagonism increased between the groups so did name calling and singing offensive songs directed at the outgroup. Following this, the groups become so hostile towards each other that they refused to eat in the same dining room together.

Next the experimenters arranged a series of joint activities for the two groups to engage in. The groups were taken to a different location and were told that they are had a shortage of drinking water due to vandalization of their water supply and the two groups had to work together repair the damage done by the vandals. During the repair significant cooperation and a reduction in antagonism was observed between the members of the two groups. During the second joint activity the groups needed to interact with each other and agree upon a movie all of them should watch. Later, during the dinner after the film the boys were eating together with no hostility. The study is so often cited to support the contact hypothesis that many undergraduate students are taught that this simple contact can remedy prejudice.

There have been a number of recent investigations of the contact hypothesis that suggest that the process is not be as straightforward as Sherif et al. (1961) described it. For instance, in one of the most supportive studies of the validity of Allport's contact hypothesis, Pettigrew and Tropp (2006) report the results of an extensive meta-analysis of studies carried out between 1940 and 2000 on intergroup contact and its reduction on prejudice. Meta-analysis is combines the results of many studies that deal with a set of related research hypotheses to determine the overall results of all the studies combined (Aron, Aron, & Coups, 2011). The logic behind the use of meta-analytic studies is that combining more studies results in more sound evidence for or against the support of a hypothesis (greater power). The resulting statistic, the effect size, for which a weighted average is often the output of meta-analyses, indicates the relative overall strength of the findings in the studies combined. Cohen's classical interpretation of effect size suggests that an effect sizes of 0.2 to 0.3 are considered a small effect, an effect size of around 0.5 a medium effect, and effect sizes from 0.8 upward are considered large effects (Cohen, 1992).

Pettigrew and Tropp report the results of 713 independent samples from 515 studies. The overall effect size was -.215, indicating a small effect, albeit it was significant due to the large number of studies examined. The more studies or subjects examined the greater the probability that the statistic will be significant (Aron, Aron, & Coups, 2011). The contact hypothesis was supported as the effect size indicates that subjects scores on measures of prejudice decreased as contact was implemented. Multiple additional tests indicated that the finding did not result from areas of bias such as publication biases, and the more the studies were controlled the more robust the mean effects were. The effects of contact appeared to generalize to the entire outgroup and contact settings. Pettigrew and Tropp also observed similar patterns of effectiveness for the contact hypothesis for ethnic targets and samples with other types of targets (e.g., obese, short, gender, etc.) suggesting that Allport's contact theory can be extended to groups other than racial groups.

Another interesting finding was that contact was significantly associated with reduced prejudice even when Allport's four pre-requisite conditions were not met (although the reduction of prejudice was greater if they were). Pettigrew and Tropp purport that these four optimal conditions should not be regarded as mutually independent but as functioning together in order to reduce prejudice and that the conditions may not be essential for prejudice reduction.

There has been an increasing interest in the use of imagery in reducing prejudice. Turner, Crisp & Lambert (2007) proposed that even imagining intergroup contact may have beneficial effects on intergroup attitudes. In three experiments, they demonstrated that thinking about contact with members of an outgroup improved attitudes toward the outgroup as a whole and reduced perceptions of outgroup homogeneity (that members of the outgroup all share the same undesirable traits). In two of their experiments participants were asked to imagine a conversation with an elderly person and in a third experiment participants imagined they were sitting next to a homosexual person in the train and conversing with them until they reached their stop. Under all three contact conditions participants reported significantly less intergroup anxiety and more positive attitudes toward the outgroup.

Following up on previous studies of imagery and contact, Stathis and Crisp (2008) completed three studies that investigated the conditions under which imagining intergroup contact would lead to greater projection of positive traits to outgroups. Two imagined contact conditions were compared: a positive imagined interaction vs. A neutral contact experience, because previous research demonstrated the effective role of positive contact rather than neutral or simple quantitative contact in improving intergroup relations. Contrast analyses were utilized as opposed to the use of omnibus statistics such as ANOVA. In the first experiment imagined contact predicted greater self-outgroup positive trait overlap for majority group members but not minority ethnic groups. In the second experiment imagined contact led to greater projection of positive traits to the outgroup for people who do not identify strongly with their own group compared to those that have higher identification with the ingroup. In the third experiment imagined contact led to greater projection of positive traits to the outgroup when a person's self-image was prominent compared to when the outgroup image was prominent.

The question asked by some is whether contact reduces prejudice or whether prejudice reduces contact. Binder, Zagefka, Brown, Funke et al., (2009) conducted a longitudinal field survey with secondary school students from Belgium, England, and Germany. The participants included a large number of subjects of ethnic minorities and ethnic majorities. The results were analyzed in terms of two indicators for prejudicial attitudes: negative intergroup emotions and desire for social distance. A series of path analyses was performed. Path analyses examine the correlations between variables to produce a predicted pattern of causal relations (Aron, Aron, & Coups, 2011). The results indicated a bidirectional mode of effects; both contact effects and prejudice effects were found. Contact reduced prejudice (but these reductions were not significant in minority participants). Furthermore, attitudes of prejudice also reduced contact for both groups. Across all subjects the effects of contact on negative emotions toward the outgroup were more robust when outgroup contacts were perceived as being stereotypical of their group.

The effects of contact were also mediated by intergroup anxiety, the more anxious one was about contact with the outgroup the less effects contact produced.

There is recent research that has explored the effects of intergroup contact on implicit forms of bias. Implicit attitudes are often measured by physiological reactions such as galvanic skin responses and others that reflect autonomic nervous system reactions. These reactions reflect conditioned associations and can be classified. For example, Turner, Hewstone, and Voci (2007) assessed the effects of contact on implicit and explicit attitudes toward elderly persons or toward Asians. Interestingly the overall amount of intergroup contact such as the proportion of one's neighbors who are outgroup members were found to be better predictors of lower implicit prejudice than measures of the quality of contact (which predicted explicit attitudes). Moreover, the effects of contact on implicit attitudes was not mediated by factors that typically mediate explicit attitudes, but demonstrated a positive impact on implicit attitudes. This suggests that simple exposure to the outgroup may be valuable in reducing unconscious biases. These findings also support the notion that implicit and explicit attitudes are qualitatively distinct from one another and do not reflect different components of the same attitude. Thus, they need to both be considered when developing interventions aimed at combating prejudice.

Despite an extensive body of evidence to support of the contact hypothesis, some have complained that too much of the research has been performed in the "experimental" than the "real world" context. Some researchers report that there are situations where contact can result in increased prejudice. For example in the real world thee are often cultural or other related reasons why contact does not reduce prejudice in people. Guerin and Guerin (2007) outline numerous subtle discriminations that exist in refugee communities during outgroup contacts and that would not be measured by the type of cross-sectional research performed in the social sciences. In a similar vein, Putnam (2007) found that in the short-term immigration and exposure to ethnic diversity reduces intergroup camaraderie and trust between interracial groups. There are areas where this friction between groups has been present for a long time the traditions of these groups make it very difficult for mere contact being able to reduce prejudice. Putnam notes that people in these areas retreat and dismiss contact with the outgroup as undesirable.

Findings

There are several major findings that the research on the contact hypothesis has produced. First, the bulk of the research suggests that contact between groups has an effect on reducing prejudice. The famous Robber's Cave study provides an early qualitative example of how contact, under certain conditions, can reduce prejudicial attitudes in young people. In the meta-analysis by Pettigrew and Tropp (2006) a large number of studies were examined and the overall effect size indicated that there was an effect of contact on the reduction of prejudice, regardless of whether the four conditions originally specified by Allport were present or not. However, the reduction in prejudice was more robust as the conditions were satisfied.

Secondly, consistent with many forms of psychotherapy that use imagery to reduce anxiety and phobias (e.g., Wolpe, 1958), there is an emerging body of research that indicates that the use of imagery can be applied to the contact hypothesis may also reduce prejudice. This finding suggests that actual physical contact may not be required to reduce prejudicial attitudes in people.

Third, there is research to support the notion that contact can affect implicit and explicit attitudes towards outgroup members. The research suggests that contact works in different ways on explicit and implicit attitudes; however, there are real changes in prejudicial attitudes from contact with outgroup members and the effects are not just stated to be politically correct but are real changes.

Fourth, although originally targeted at racial prejudice, the contact hypothesis has been shown to be effective with other types of outgroup prejudice such as attitudes towards the elderly, obese, homosexuals, or other groups. This results in a broad array of applications.

And finally, there is research to suggest that under some conditions contact may not be effective in reducing prejudice or may be less effective. In the path analyses by Binder et al. (2009) and the imagery study by Stathis and Crisp (2008) it was observed that contact had less of an effect on minority subjects than majority subjects. The imagery studies of Stathis and Crisp (2008) also indicated that when one had less of an identity with their own group or that one was less focused on outgroup differences imaginary contact was more effective. Putnam (2007) and Guerin (2007) have outlined "real world" situations where contact may have a negative effect on reducing prejudice.

Discussion

In general it appears that the literature supports the notion that the contact hypothesis does offer some validity in reducing prejudicial attitudes. The research has progressed a long way from the Sherif et al. (1961) Robber's Cave study which has many methodological concerns such as poor measurement of the dependent variable (prejudice) and many potential confounds (was it the cooperation of the groups or mere exposure that resulted in the behavior change).